Startup Idea: Automated Fraudulent Transaction Analysis System

Startup idea to develop intelligent software that can analyze transactions flagged by a card processor's system for potential fraud, identify trends and patterns, dismiss false positives and only retain truly suspicious transactions for manual review. The goal is to replace the current manual process that uses an excel spreadsheet.

I work for a card processor and we have an issue where we set up certain rules to monitor transactions and when one of the rules gets triggered, information about that transaction is added to a daily report. I then have to manually look through the transactions on that daily report and figure out whether they are fraudulent or not. Software that could analyze those triggered transactions would be helpful. Maybe it could analyze trends and patters for certain cardholders and remove false positives from the report and only keep transactions that are truly fraudulent for manual review. We were trying to build one ourselves, but we were lacking resources so this is currently tracked in and excel spreadsheet which is not very convenient. I don't know if there is a universal solution that could work with our system.